Skip to Content.
Sympa Menu

svadev - Re: [svadev] [Proposal] Change return values in SAFECode/test

svadev AT lists.siebelschool.illinois.edu

Subject: Svadev mailing list

List archive

Re: [svadev] [Proposal] Change return values in SAFECode/test


Chronological Thread 
  • From: John Criswell <criswell AT illinois.edu>
  • To: Santosh Nagarakatte <santosh.nagarakatte AT gmail.com>
  • Cc: svadev AT cs.uiuc.edu
  • Subject: Re: [svadev] [Proposal] Change return values in SAFECode/test
  • Date: Wed, 23 May 2012 14:12:01 -0500
  • List-archive: <http://lists.cs.uiuc.edu/pipermail/svadev>
  • List-id: <svadev.cs.uiuc.edu>
  • Organization: University of Illinois

On 5/23/12 12:28 PM, Santosh Nagarakatte wrote:
Hi all,

While I was experimenting with SoftBound and safecode tests, I noticed
that few of the tests return non-zero values from main (example
test/core/buffer-001.c, ...).

I am not using the llvm testing infrastructure.
I have a custom python based test infrastructure.
Having non-zero return values from main (which is the exit code)
complicates automated checking (distinguishing false violations/real
violations).

I propose that we change the return value to zero.
I understand that values were returned to prevent compiler from
optimizing the code away.
We can attain the same effect with print statements.

I agree and, IIRC, ran into some of the same problems even when using our testing infrastructure.

Santosh, please file a bug for this if you don't hear objections from anyone else. Then feel free to either fix it yourself (if you want) or let one of us fix it.

-- John T.


Thanks,
Santosh







Archive powered by MHonArc 2.6.16.

Top of Page